Add new attachment

Only authorized users are allowed to upload new attachments.

List of attachments

Kind Attachment Name Size Version Date Modified Author Change note
atr
fig-FORTH1.4S-1.atr 92.2 kB 1 05-Jul-2016 10:53 Roland B. Wassenberg fig-FORTH1.4S-1
atr
fig-FORTH1.4S-2.atr 92.2 kB 1 05-Jul-2016 10:53 Roland B. Wassenberg fig-FORTH1.4S-2

This page (revision-72) was last changed on 01-May-2018 03:30 by Roland B. Wassenberg  

This page was created on 20-Feb-2010 21:56 by Carsten Strotmann

Only authorized users are allowed to rename pages.

Only authorized users are allowed to delete pages.

Difference between version and

At line 1 changed one line
!!! Forth
!!! Forth
At line 3 added 29 lines
!Background
Forth is an interpreted programming language that uses a stack-based metaphor in an effort to reduce memory requirements as much as possible. It is perhaps the only successful example of such a language; the PostScript system was derived from it and saw much widespread use, but this was hidden inside the printers and not seen by end-users or programmers.
Stack-based languages simplify the interpreter's parser considerably because the data for an instruction always appears in the source code before the instructions that will use it. To see why this helps, consider this typical line of [Basic]:\\
\\
{{A = 10 + 20 * B}}\\
\\
To perform this line, the interpreter has to read the entire line, look up the value of B (let's say 30), realize that the * has to be performed before + and order the instructions correctly, and then finally convert those into instructions something like:\\
\\
{{get(B,temp1)}} - get the value in B and store it in temp1\\
{{multiply(20,temp1,temp2)}} - multiply that value by 20 and store the result in temp2\\
{{add(10,temp2,temp3)}} - add 10 to temp2 and store the result in temp3 \\
{{put(temp3,A)}} - store the value of temp3 into the variable A\\
\\
In contrast, in a stack-based system, the programmer organizes the code in the fashion it will ultimately be performed. The equivalent would be something like:\\
\\
{{B 20 mul}}\\
{{10 add}}\\
\\
When this code is performed, the interpreter pushes the value of B on the stack, then 20. It then encounters the mul, which removes the last two items, the 30 and 20, multiplies them, and puts the result back on the stack. Next, it pushes 10 on the stack, leaving the top two locations containing 60 and 10. It then encounters add, taking the two values, adding them, and putting the result back on the stack. The top of the stack now contains the result, 70.
Notice that the stack-based version ''has no temporary values'', and only reads a single instruction at a time, not an entire line of code. As a result, the parser is much simpler, smaller and requires less memory to run. This, in turn, generally makes it much faster, comparable to compiled programs.
Another key aspect of the language was Forth's inherently multitasking design. The program could set up separate stacks and feed different code into each one. The Forth kernel would run each of these stacks in turn, so all Forth programs had access to these features. This made writing multithreaded code very easy, so one could, for instance, have a thread reading the joystick as it moved, and then read that value in a game loop in another stack.
The downside to the stack-based approach is that it makes the language difficult to understand by mere mortals. Even tutorials purporting to show how simple it was often ended in an unreadable mess. As a result, Forth was subject to perhaps one of the longest running fanboi wars since APL was invented. Constantly derided by practically everyone in the industry, it saw some interest in spite of this, but little commercial software emerged. The singular exception is the PostScript system, which is essentially a version of Forth modified to produce graphics output.
The Atari 8-bits were being sold right in the middle of this battle, and as a result there was a fair amount of support on the platform and some interest in the press.
At line 63 changed one line
* [SNAUT] (probably vaporware)
* [SNAUT]
At line 79 removed one line
Version Date Modified Size Author Changes ... Change note
72 01-May-2018 03:30 6.695 kB Roland B. Wassenberg to previous
71 28-Apr-2018 15:44 6.695 kB Maury Markowitz to previous | to last
70 28-Apr-2018 15:38 6.295 kB Maury Markowitz to previous | to last
69 28-Apr-2018 15:36 6.27 kB Maury Markowitz to previous | to last
68 26-Apr-2018 22:31 4.923 kB Maury Markowitz to previous | to last
67 26-Apr-2018 22:19 4.336 kB Maury Markowitz to previous | to last
66 29-Aug-2017 04:50 3.271 kB Roland B. Wassenberg to previous | to last
65 25-Feb-2017 16:55 3.292 kB Carsten Strotmann to previous | to last
64 05-Jul-2016 10:56 3.153 kB Roland B. Wassenberg to previous | to last
63 02-Jul-2016 01:07 3.103 kB Roland B. Wassenberg to previous | to last
62 01-Jul-2016 18:40 3.062 kB Roland B. Wassenberg to previous | to last
61 23-Jan-2014 18:28 3.048 kB Carsten Strotmann to previous | to last
« This page (revision-72) was last changed on 01-May-2018 03:30 by Roland B. Wassenberg